Image segmentation of topographical objects employing illumination parameter variations and correlation for quality control applications

Abstract

3-D topography of objects is often obtained only on a qualitative basis by a sequential reconstruction process using segmented images obtained under different illumination conditions. The approach we present in this paper is based on the direct segmentation and topography analysis of a topographic scene using a sequence of images obtained by varying the illumination parameters in a pre-selected format. Thereby a correlation of the grey level variations of the individual pixels with the variations of the illumination parameters is employed to achieve superior segmentation and topographical information. An advantage is that the pipeline of separate operations viz. image fusion, segmentation topographical analysis is reduced to just one direct operation. The method has been successfully applied to determine the topography of grinding surfaces in machining in manufacturing processes.